CN107358138A - Bearing calibration, mobile terminal and the storage device of nonlinear distortion EAN bar codes - Google Patents

Abstract

The invention discloses a method for correcting nonlinearly distorted EAN barcodes, a mobile terminal and a storage device. Wherein, the correction method includes: step S1, sequentially performing binarization processing and inversion processing on the non-linearly distorted EAN barcode image to obtain an inverse image including the background area and the foreground area; Screen and extract the stripes, and then use the skeleton extraction technology to refine the retained stripes to obtain a set of single-pixel thinned lines; step S3, perform parallel correction on the adjacent single-pixel thinned lines in sequence , to obtain a set of parallel thinning lines; step S4 , perform area correction on the parallel thinning lines by using block sampling and bilinear interpolation technology, and obtain a corrected image of the EAN barcode with the same size. The EAN barcode correction method of the present invention adopts block sampling and interpolation technology to correct step by step, has high stability and precision, and small calculation amount.

Description

Correction method, mobile terminal and storage device for non-linearly distorted EAN barcode

technical field

The invention relates to the field of graphics and image processing, in particular to a method for correcting nonlinearly distorted EAN barcodes, a mobile terminal and a storage device.

Background technique

EAN (European Article Number Bar Code) is a kind of barcode for commodities developed by the International Article Numbering Association, which has been widely used in many fields such as commodity circulation, book management, transportation, medical and health care, and finance and banking. In the actual application environment, the barcode captured by the camera is often affected by the shooting angle, the deformation of the barcode attachment surface, etc., which causes the quality of the barcode image to degrade, thereby seriously reducing the accuracy of barcode recognition.

The so-called EAN barcode distortion correction technology is a process of correcting nonlinear geometric distortions such as angle rotation and local distortion after separating the barcode area.

There are two main types of existing barcode distortion correction technologies: (1) barcode correction methods based on line detection. With the help of straight line detection methods such as Hough transform, the spatial distribution statistical characteristics of black and white stripes of barcodes are used to correct linear distortion barcodes such as rotation deformation or affine deformation. (2) Barcode correction method based on geometric space transformation. The distorted barcode is corrected by collecting the information of several control point pairs and solving the solution of the equation system, and then using the space transformation matrix.

The existing barcode correction methods have the following disadvantages: (1) Line detection methods can only deal with linear distortions such as rigid body transformations or affine transformations introduced by different shooting angles, but cannot deal with nonlinear distortions such as local distortions, and their application range is limited. (2) The geometric space transformation method uses the matrix operation method to solve the equation system, and the calculation amount of the solution is large and complicated. In some specific environments, the equation system itself may have no solution or infinitely many solutions, resulting in low reliability and poor real-time performance.

Therefore, the prior art still needs to be improved and developed.

Contents of the invention

In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a correction method for nonlinearly distorted EAN barcodes, a mobile terminal and a storage device, aiming at solving the problem that the prior art cannot handle nonlinear distortion such as local distortion and low reliability , The problem of poor real-time performance.

Technical scheme of the present invention is as follows:

A method for correcting nonlinearly distorted EAN barcodes, comprising:

Step S1, sequentially performing binarization processing and inversion processing on the non-linearly distorted EAN barcode image to obtain an inversion image including a background area and a foreground area;

Step S2: Filter and extract the stripes in the foreground area according to specific screening rules, and then use the skeleton extraction technology to refine the retained stripes to obtain a single-pixel thinned line set;

Step S3. Parallel correction is performed on adjacent single-pixel thinning lines in order to obtain a set of parallel thinning lines;

Step S4, using sub-block sampling and bilinear interpolation techniques to perform area correction on the parallel thinning lines to obtain a corrected image of the EAN barcode with the same size.

The method for correcting the nonlinearly distorted EAN barcode, wherein, in the step S1, the binarization processing method is a global binarization algorithm or a local adaptive binarization algorithm.

The method for correcting the nonlinearly distorted EAN barcode, wherein, in the step S1, the process of the inverse color processing is specifically: the light and dark stripes in the binarized image correspond to pixel values 0 and 255 respectively, and the pixel value 0 corresponds to The area with pixel value 255 is called the foreground area.

The correction method of the nonlinear distortion EAN barcode, wherein, in the step S2, the screening rule is: calculate the pixel area of each foreground area, if the area of a single area is greater than the average value of the area of all foreground areas, then keep the foreground region, otherwise delete it.

The method for correcting the nonlinearly distorted EAN barcode, wherein the step S3 specifically includes:

Step S31, select two adjacent single-pixel thinning lines in order, and record them as L _left and L _right respectively. Use the Hausdorff distance measurement method to calculate the statistical distance between the single-pixel thinning lines: Among them, l ₁ and l ₂ represent the points on L _left and L _right respectively, and d(*,*) represents the two-dimensional Euclidean distance;

Step S32, draw the thinned line of single pixel into a straight line segment, and the distance between adjacent straight line segments is D(L _left , L _right ) described in step S31;

Step S33 , repeating step S31 and step S32 , performing parallel correction on the thinning fringes of the remaining single pixels in order to obtain a set of parallel thinning fringes.

The method for correcting the nonlinearly distorted EAN barcode, wherein the step S4 specifically includes:

Step S41, connect the corresponding sampling points on the two adjacent parallel thinning lines one by one, and use bilinear interpolation technology to map the pixels on the connection line to the corresponding connection line between the corresponding parallel straight line segments;

Step S42: Traversing and repeating the above process pixel by pixel from top to bottom, correcting the pixels at all positions between the two adjacent thinning lines, and then traversing all the adjacent thinning lines in order to complete all the position pixels between the middle lines correction;

Step S43 , performing single-line sampling area correction on the starting thinning line and the ending thinning line.

The method for correcting the nonlinearly distorted EAN barcode, wherein, in the step S41, the corresponding point mapping relationship between the pixels on the connection line and the parallel straight line segments is:

Where (x _left , y _left ) is the coordinates of point P _left on the thin line on the left, (x _right , y _right ) is the coordinates of the corresponding point P _right on the thin line on the right, (x _S , y _S ) is the coordinates of P _left and The coordinates of a certain pixel point on the P _right line, (x, y) are the coordinates of the corresponding points between the parallel straight line segments.

The correction method of the nonlinear distortion EAN barcode, wherein, the step S43 is specifically: calculating the normal direction of the sampling point on the edge line, calculating the Euclidean distance between the intersection point of the normal line and the image boundary and the sampling point, Then extend the above-mentioned Euclidean distance horizontally to the image boundary direction on the point corresponding to the sampling point on the straight line segment to obtain the corresponding mapping boundary point. In this way, the area correction of the start thinning line and the ending thinning line is completed.

A mobile terminal, comprising a processor and a memory connected in communication with the processor, the memory stores a computer program, and when the computer program is executed by the processor, the correction of the non-linear distortion EAN barcode as described above is realized Method; the processor is used to call the computer program in the memory to execute any one of the methods for correcting the nonlinearly distorted EAN barcode described above.

A storage device, the storage device stores a computer program, and the computer program can load and execute the method for correcting nonlinearly distorted EAN barcodes as described above.

Beneficial effects: the present invention provides a method for correcting nonlinearly distorted EAN barcodes, a mobile terminal and a storage device. The method for correcting EAN barcodes of the present invention is based on the extraction of the shape features of the barcode itself, without any restrictions on the type of distortion, and can Handle any nonlinear distortion and deformation with high versatility. In addition, the present invention adopts block sampling and interpolation technology to correct step by step, so the stability and precision are high, and the amount of calculation is small.

Description of drawings

FIG. 1 is a flowchart of a method for correcting nonlinearly distorted EAN barcodes of the present invention.

Fig. 2a is a schematic diagram of points taken on adjacent thinning lines in the adjacent line area correction step of the present invention.

Fig. 2b is a schematic diagram of corresponding points on parallel line segments in the correction step of the adjacent line area of the present invention.

Fig. 3a is a schematic diagram of calculating the normal direction of the curve in the single-line area correction step of the present invention.

Fig. 3b is a schematic diagram of the mapping of sampling points to the calibrated figure in the single-line region correction step of the present invention.

FIG. 4 is a functional block diagram of a mobile terminal based on a nonlinear distortion EAN barcode correction method according to the present invention.

FIG. 5 is an effect diagram of the non-linearly distorted EAN barcode image after binarization and color inversion processing in the method for correcting the non-linearly distorted EAN barcode of the present invention.

FIG. 6 is an effect diagram after filtering and thinning the nonlinearly distorted EAN barcode image in the nonlinearly distorted EAN barcode correction method of the present invention.

FIG. 7 is an effect diagram of correction results of adjacent thinning line regions in the nonlinear distortion EAN barcode correction method of the present invention.

FIG. 8 is an effect diagram of the final correction result of the nonlinear distortion EAN barcode correction method of the present invention.

detailed description

The present invention provides a method for correcting non-linearly distorted EAN barcodes, a mobile terminal and a storage device. In order to make the purpose, technical solutions and effects of the present invention clearer and clearer, the present invention will be further described in detail below. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

The present invention provides a preferred embodiment of a correction method for nonlinearly distorted EAN barcodes. The correction process is shown in Figure 1. The correction steps include:

Step S1, sequentially perform binarization processing and inversion processing on the non-linearly distorted EAN barcode image to obtain an inversion image including a background area and a foreground area.

Specifically, the barcode image is binarized first. According to the acquisition quality of the actual barcode image, the binarization algorithm can choose a global binarization algorithm such as fixed threshold method or Otsu method or a local adaptive binarization algorithm. Then, reverse color processing is performed on the obtained binarized image. After inversion, the original light and dark stripes correspond to pixel values 0 and 255, respectively. Generally, the area corresponding to pixel 0 is called the background area, and the area corresponding to pixel 255 is called the foreground area.

Step S2: Filter and extract the stripes in the foreground area according to specific screening rules, and then use the skeleton extraction technology to refine the retained stripes to obtain a set of thinned single-pixel lines.

Specifically, Freeman chain code technology is used to extract the boundary contour of the foreground area, and the pixel area of each isolated foreground area (denoted as S _i , i=1, 2, 3,...) and the average pixel area of all foreground areas ( denoted as S ₀ ), if S _i ≥ S ₀ , keep the foreground area, otherwise delete it. In this way, the streak screening step is completed.

From the above screening rules, it can be seen that the set of reserved foreground areas are all stripes with relatively large area and non-single-pixel width. Therefore, the skeleton extraction technology is used to refine the foreground area, and finally a set of lines with a single-pixel width is obtained.

It should be noted that the present invention does not extract the narrow stripes corresponding to the foreground area that is smaller than the average pixel area of all foreground areas, in order to prevent large errors in the extraction process and affect the final correction accuracy. For correction, the traditional rigid body transformation or affine transformation and other processing methods for linear distortion cannot process wide stripes and thin stripes in different ways.

Step S3 , parallel correction is performed on adjacent single-pixel thinning lines in order to obtain a set of parallel thinning lines.

After obtaining the single-pixel thinning line set, additionally create a set of straight line segments with the same size as the single-pixel thinning stripes. Specifically, two adjacent thinning lines are selected in sequence (from left to right or from right to left, "sequence" in the present invention should be understood in this way), which are respectively marked as L _left and L _right . Due to the influence of nonlinear distortion, the lines L _left and L _right are no longer parallel line segments. Calculate the statistical distance between lines using the Hausdorff distance metric:

Among them, l ₁ and l ₂ represent points on the lines L _left and L _right respectively. d(*,*) represents the two-dimensional Euclidean distance.

Assuming that the horizontal direction is the x direction and the vertical direction is the y direction, draw the corrected vertical line segments CL _left and CL _right according to the following rules. The following uses CL _left as an example to illustrate:

In the first step, the horizontal direction coordinate of CL _left is defined as the average value of horizontal direction x-coordinates of all points in the line L _left . The vertical y-coordinate of its upper endpoint is defined as the vertical y-coordinate of the upper endpoint of the line L _left .

In the second step, the vertical y-coordinate value of the lower endpoint is defined as the sum of the vertical y-coordinate value of the upper endpoint of CL _left and the pixel length of the line L _left (the total number of 255 pixels).

In the third step, the pixel values of the CL _left line segment are assigned values from the pixel values of L _left in order from top to bottom.

The horizontal coordinate of CL _right is defined as the horizontal x coordinate of CL _left plus D(L _left , L _right ). The definition of other coordinate information is exactly the same as CL _left .

According to the above steps, the single-pixel thinning lines are drawn into straight lines of the same length, and all the thinning lines are corrected into parallel thinning lines in sequence, and the distance between adjacent parallel lines is D( L _left , L _right ).

Step S4, using sub-block sampling and bilinear interpolation techniques to perform area correction on the parallel thinning lines to obtain a corrected image of the EAN barcode with the same size.

First perform double-line area correction, that is, take corresponding sampling points on two adjacent thinning lines, and connect the sampling points one by one, for example, the first pixel on L _left and the first pixel on L _right Pixel connection, the connection between the second pixel on L _left and the second pixel on L _right , the connection between the third pixel on L _left and the third pixel on L _right , and so on. The bilinear interpolation technology is used to map the pixel points on the connection line to the corresponding connection line between the corresponding parallel straight line segments.

Specifically, for example, starting from the left, select two adjacent thinning lines L _left and L _right , and set the corresponding corrected parallel straight line segments as CL _left and CL _right . The lengths of L _left and CL _left are equal, and the lengths of L _right and CL _right are equal, which are recorded as Length _left and Length _right respectively. Assuming that P _left is the kth coordinate point sorted from top to bottom on L _left , then its corresponding point P _right on L _right is defined as sorted from top to bottom as coordinate point. The corresponding points CP _left and CP _right of P _left and P _right on the corrected image are defined as k and coordinate point. As shown in Figure 2a and Figure 2b.

The process of area correction between adjacent thinning lines is as follows:

Assume that the coordinates of P _left are (x _left , y _left ), the coordinates of P _right are (x _right , y _right ), the coordinates of CP _left are (cx _left , cy _left ), and the coordinates of CP _right are (cx _right , cy _right ). For any point CP(x,y) on the line segment CP _left CP _right , its corresponding coordinate P(x _S , y _S ) on the original image is calculated by the following formula:

Then the pixel value of CP(x,y) is assigned as the pixel value of P(x _S ,y _S ).

By traversing and repeating the above process pixel by pixel from top to bottom, pixels at all positions between two adjacent line regions can be corrected. In this way, all adjacent lines are traversed in sequence, and the correction of all position pixels between the lines can be completed.

The above correction methods are all aimed at the correction of the middle area, including not only the correction of the extracted fringes, but also the direct correction of the unextracted thin fringes. However, the area to the left of the leftmost line and the area to the right of the most right line where the start line and the end line are located are not included, and a single-line sampling area correction needs to be performed on the two thinning lines on the edge. The correction method is to calculate the normal direction of the sampling point on the edge line, calculate the Euclidean distance between the intersection point of the normal line and the image boundary and the sampling point, and then extend the point corresponding to the sampling point on the straight line segment horizontally toward the image boundary direction The above-mentioned Euclidean distance is to obtain the corresponding mapping boundary points, and the area correction of the starting thinning line and the ending thinning line can be completed by this method.

Taking the rightmost line as an example, the method for correcting the sampling area of a single line is described below.

Unlike double lines, single lines do not have a corresponding edge on the right. As defined in the preceding steps, it is assumed that the coordinates of P _left are (x _left , y _left ), and the coordinates of CP _left are (cx _left , cy _left ). The method of coordinates of the virtual P _right is given below, as shown in Fig. 3a and Fig. 3b.

First calculate the normal direction at the point P _left . The specific method for calculating the normal direction of the curve is: the given curve is sorted in ascending order according to the size of the y coordinate, which is recorded as {Curve[i],i=1,2,..., N}. Any point P on the curve, with the serial number p, can be divided into two cases:

(1) The point P falls in the head or tail area of the curve, that is, the sequence number of P is p≤5 or p≥N-4. Taking the point P falling on the head as an example, the sampling point set Curve[0], Curve[1],...,Curve[p+5] is taken. Fit the set data to a straight line fitL, take the direction of fitL as the tangent direction of the point, and then calculate its normal direction to the right as the approximate normal direction of point P.

(2) Point P falls in the middle of the curve. The collection of sampling points is Curve[p-5],...,Curve[p],...,Curve[p+5]. The remaining calculation method is the same as step (1). I won't repeat them here.

Then a ray is taken along the normal direction, and the intersection point of the ray and the image boundary is marked as P _right . Calculate its Euclidean distance Length _left . Since CL _left itself is a vertical line segment, the value of CP _right is the coordinate of CP _left extending to the right by Length _left . So far, P _left , P _right and their corresponding CP _left and CP _right have been obtained. Following this calculation process, the area to the right of the rightmost line can be corrected. In the same way, the area where the leftmost line goes to the left can be corrected.

The present invention also provides a mobile terminal. As shown in FIG. 4 , the mobile terminal includes: a processor (processor) 10 , a memory (memory) 20 , a communication interface (Communications Interface) 30 and a bus 40 . Wherein, the processor 10, memory 20, and communication interface 30 complete mutual communication through the bus 40; the communication interface 30 is used for information transmission between communication devices of the mobile terminal; the processor 10 It is used to call the computer program in the memory 20 to execute the methods provided by the above method embodiments.

When the system starts, the mobile terminal obtains the non-linearly distorted EAN barcode image, and sequentially performs binarization and inversion processing to obtain an inversion image including the background area and the foreground area. The processed effect is shown in Figure 5; then Filter and extract the stripes in the foreground area according to specific screening rules, and use the skeleton extraction technology to refine the retained stripes to obtain a single-pixel thinned line set, as shown in Figure 6; and then compare the adjacent Single-pixel thinning lines are parallel-corrected in pairs in order to obtain a set of parallel thinning lines. Finally, block sampling and bilinear interpolation techniques are used to perform area correction on parallel thinning lines to obtain EAN barcode corrections of the same size. Images, as shown in Figure 7 and Figure 8.

The present invention also provides a storage device, wherein the storage device stores a computer program that can be executed to implement the method for correcting the nonlinearly distorted EAN barcode.

In summary, the present invention provides a method for correcting nonlinearly distorted EAN barcodes, a mobile terminal and a storage device. The method for correcting EAN barcodes of the present invention is based on the extraction of the shape features of the barcode itself, without any restrictions on the type of distortion , can handle any nonlinear distortion deformation, and has high versatility. In addition, the present invention adopts block sampling and interpolation technology to correct step by step, so the stability and precision are high, and the amount of calculation is small.

Of course, those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware (such as processors, controllers, etc.), and the programs can be stored in a In the computer-readable storage medium, the program may include the processes of the above-mentioned method embodiments when executed. The storage medium mentioned herein may be a memory, a magnetic disk, an optical disk, and the like.

It should be understood that the application of the present invention is not limited to the above examples, and those skilled in the art can make improvements or transformations according to the above descriptions, and all these improvements and transformations should belong to the protection scope of the appended claims of the present invention.

Claims (10)

1. A correction method for nonlinear distortion EAN barcode, characterized in that, comprising: Step S1, sequentially performing binarization processing and inversion processing on the non-linearly distorted EAN barcode image to obtain an inversion image including a background area and a foreground area; Step S2: Filter and extract the stripes in the foreground area according to specific screening rules, and then use the skeleton extraction technology to refine the retained stripes to obtain a single-pixel thinned line set; Step S3. Parallel correction is performed on adjacent single-pixel thinning lines in order to obtain a set of parallel thinning lines; Step S4, using sub-block sampling and bilinear interpolation techniques to perform area correction on the parallel thinning lines to obtain a corrected image of the EAN barcode with the same size. 2. The correction method of nonlinear distortion EAN barcode according to claim 1, characterized in that, in said step S1, the method of said binarization processing is a global binarization algorithm or a local adaptive binarization algorithm . 3. The method for correcting nonlinearly distorted EAN barcodes according to claim 1, characterized in that, in the step S1, the process of the inverse color processing is specifically: the light and dark stripes in the binarized image correspond to pixel values respectively 0 and 255, the area corresponding to the pixel value 0 is called the background area, and the area corresponding to the pixel value 255 is called the foreground area. 4. The method for correcting nonlinearly distorted EAN barcodes according to claim 1, wherein in said step S2, said screening rule is: calculate the pixel area of each foreground area, if the area of a single area is larger than all foreground areas If it is the average value of the area, the foreground area is retained, otherwise it is deleted. 5. the correction method of nonlinear distortion EAN barcode according to claim 1, is characterized in that, described step S3 specifically comprises: Step S31, select two adjacent single-pixel thinning lines in order, and record them as L _left and L _right respectively. Use the Hausdorff distance measurement method to calculate the statistical distance between the single-pixel thinning lines: Among them, l ₁ and l ₂ represent the points on L _left and L _right respectively, and d(*,*) represents the two-dimensional Euclidean distance; Step S32, draw the thinned line of single pixel into a straight line segment, and the distance between adjacent straight line segments is D(L _left , L _right ) described in step S31; Step S33 , repeating step S31 and step S32 , performing parallel correction on the thinning fringes of the remaining single pixels in order to obtain a set of parallel thinning fringes. 6. the correction method of nonlinear distortion EAN barcode according to claim 1, is characterized in that, described step S4 specifically comprises: Step S41, connect the corresponding sampling points on the two adjacent parallel thinning lines one by one, and use bilinear interpolation technology to map the pixels on the connection line to the corresponding connection line between the corresponding parallel straight line segments; Step S42: Traversing and repeating the above process pixel by pixel from top to bottom, correcting the pixels at all positions between the two adjacent thinning lines, and then traversing all the adjacent thinning lines in order to complete all the position pixels between the middle lines correction; Step S43 , performing single-line sampling area correction on the starting thinning line and the ending thinning line. 7. The method for correcting nonlinearly distorted EAN barcodes according to claim 6, characterized in that, in the step S41, the corresponding point mapping relationship between the pixels on the connection line and the parallel straight line segments is: Where (x _left , y _left ) is the coordinates of point P _left on the thin line on the left, (x _right , y _right ) is the coordinates of the corresponding point P _right on the thin line on the right, (x _S , y _S ) is the coordinates of P _left and The coordinates of a certain pixel point on the P _right line, (x, y) are the coordinates of the corresponding points between the parallel straight line segments. 8. The method for correcting nonlinearly distorted EAN barcodes according to claim 6, wherein said step S43 is specifically: calculating the normal direction of the sampling point on the edge line, calculating the intersection point and sampling point of the normal line and the image boundary The Euclidean distance between the points, and then extend the above-mentioned Euclidean distance horizontally to the image boundary direction on the point corresponding to the sampling point on the straight line segment, that is, to obtain the corresponding mapping boundary point, and complete the start thinning line and end thinning line in this way. area correction of the lines. 9. A mobile terminal, characterized in that it comprises a processor and a memory connected in communication with the processor, the memory stores a computer program, and when the computer program is executed by the processor, any one of claims 1-8 is realized. The method for correcting the nonlinearly distorted EAN barcode; the processor is used to call the computer program in the memory to execute the method for correcting the nonlinearly distorted EAN barcode according to any one of claims 1-8. 10. A storage device, characterized in that the storage device stores a computer program capable of loading and executing the method for correcting the non-linearly distorted EAN barcode according to any one of claims 1-8.